PEOi'ESSOE OWEN ON CNEMIOKNIS, 265 



steiTial groove. The length of the bone which includes the beginning of this expansion 

 is 3 inches 6 lines ; the entire bone would be about 4 inches 6 lines in length. The 

 extreme breadth at the middle of the shaft is 4^ lines. It is thus weaker and more 

 slender than in Cereopsis, and longer in proportion to its sternal breadth than in 

 Tachyeres. It also differs from the coracoid in these and other Lamellirostrals in 

 the very slight production of the tuberosity c in advance of that (5) supporting the 

 the articular surface (a) for the humerus. The tuberosity c is divided from J by a 

 shallow groove [d) of less than half the width of the homologous one in Cereopsis ; and 

 the tuberosity b is not present in Cereopsis, or is represented (as in fig. 8, h) only by 

 the produced margin of the relatively larger and deeper facet for the humerus. The 

 process (e) joining the median facet of the scapular articular expansion is more produced, 

 more terminally expanded, both lengthwise and transversely; the latter expansion 

 inclines, as a curved lamella, toward the inner or anterior division of the tuberosity c, 

 in advance of the humeral joint. 



From the low scapular process in Cereopsis (PI. XXXVII. iig. 8, e) a ridge of bone 

 (ib.y) extends down to the middle of the coracoid, where it blends with the mesial 

 border, leaving a narrow oblong interspace, 4 lines in length, near that border. This 

 character is not present in the coracoid of Tachyeres. Such a vacuity (ib. fig. 5,_/') 

 exists in the coracoid of Cnemiornis ; but its filamentary boundary is not continued 

 from the scapular process' (e, e') ; it forms part, or is a continuation, of the sharp 

 mesial border of the shaft of the bone ; and the vacuity is a perforation of such border. 



An intermuscular ridge (ib. fig. 5, g) is continued in Cnemiornis more directly from 

 the tuberosity (<?), but sooner subsides upon the shaft than in Cereopsis ; it is resumed 

 at the lower third of the shaft, but nearer the lateral border, and bounds the fore part 

 of a flat, roughish, elongate tract, which has a continuation of the lateral border (ib. 

 fig. 7,y) for its hinder boundary. Above this tract, the shaft of the coracoid is thicker 

 in Cnemiornis than in Cereopsis and other Anserines. The hind surface of the sternal 

 half of the coracoid is feebly concave ; the sternal articular expanded end has been 

 broken away in my specimen. 



Although this coracoid is more slender, in proportion to its length, than in. Cereopsis, 

 it is thicker, and less flattened from before backward toward the sternal expansion. 

 This proportion is still more characteristic of the coracoid of Cnemiornis, in comparison 

 with that of Tachyeres, in which the whole shaft is more flattened than in Cereop)sis. 



The strength of the bone in Cnemiornis relates to its office in depressing the sternum 

 in the respiratory movements of the bird. 



In describing the humerus ' forming part of the collection of bones including a skull 

 of Binornis robustus and part of one of Aptornis, together with the tibia and other bones 

 on which was founded the genus Cnemiornis, I stated that, " from the feeble develop- 

 ment of its proximal processes," such humerus " had evidently belonged to some such 



' Zool. Trans, vol. T. pp. 396-399. 



